Multidimensional Digital Smoothing Filters for Target Detection

TL;DR

This paper introduces multidimensional digital smoothing filters designed to enhance target detection in surveillance by reducing clutter and interference while accumulating target signal power, using advanced mathematical design techniques.

Contribution

It develops a novel design methodology for recursive multidimensional filters employing discrete Laguerre polynomials and least-squares regression, linking filter characteristics to difference-equation coefficients.

Findings

Filters effectively negate correlated clutter and interference.

Simulation verifies theoretical expressions for filter parameters.

Filters improve detection of dim targets in complex backgrounds.

Abstract

Recursive, causal and non-causal, multidimensional digital filters, with infinite impulse responses and maximally flat magnitude and delay responses in the low-frequency region, are designed to negate correlated clutter and interference in the background and to accumulate power due to dim targets in the foreground of a surveillance sensor. Expressions relating mean impulse-response duration, frequency selectivity and group delay, to low-order linear-difference-equation coefficients are derived using discrete Laguerre polynomials and discounted least-squares regression, then verified through simulation.